The single generalist agent is gone. BigPanda has now formally announced its rebuilt AI Incident Assistant: trailed in May, deployed in the July release, unveiled in full on 14 August.
The shift underneath: fixed, hardcoded action plans give way to an Orchestrator that reads each request and assembles the answer from composable capabilities: Skills, Agents, Workflows and Actions. Haydn Ashcroft covered the release mechanics last month, so here is what the announcement adds, and what it doesn't change.
On complex incidents, Biggy agents chase multiple hypotheses in parallel across your estate instead of working one thread at a time. This shipped quietly in July under custom agents, where you can mark your own agents as swarm-eligible so they contribute their specialism to these investigations. Worth revisiting your agent configuration with that in mind.
Biggy narrates what it is checking and why, streaming partial findings you can act on before the investigation completes. Trust in agentic operations is earned through transparency and reversibility, not autonomy. July's approval gates delivered the reversibility. This is the transparency.
There is no new SKU and, as above, no new capability to wait for. Existing customers get the full experience at their next login, with one exception: Adaptive Memory needs enablement and configuration, so that is a conversation with your account team rather than a toggle. It is a conversation worth having: memory captures facts, proven approaches, learned mistakes and full episodes from real incidents, and recalls them on every request, so Biggy arrives at the next incident already knowing what caused the last. But look first at what will feed it. Biggy reasons over whatever your integrations and correlation supply, and a memory that compounds will compound bad foundations as readily as good ones.